For once, there have been no formal changes in staff this quarter, and the composition of the group remains as before, ie
The present staff in the group are:
| D A Byfield | Secretary | (DAB) |
| M F Chiu | Attached to FR80 | (MFC) |
| J E Crow | (JEC) | |
| M Davies | (MD) | |
| M Elder | (ME) | |
| E M Gill | (EMG) | |
| M F Guest | (MFG) | |
| P Kent | (PK) | |
| J W E Lewis | ½ attached to FR80 | (JWEL) |
| P A Machin | (PAM) | |
| V R Saunders | (VRS) | |
| P F Smith | (PFS) | |
| B Stokoe | (BS) | |
| G Symes | (GS) | |
| A J H Walter | (AJHW) |
There has been some unsettlement in the group with the impending possible reorganisation; and some people are beginning to be in some sense assigned to projects outside the formal remit of Applications Software.
The configuration selector for the CI package is now operational, and hopefully this completes the programming effort required for the system. The selector provides a mechanism for selecting a tractable number of configurations in a systematic fashion from a list that is formally unmanageable. Two variants have been coded and tested; in one case each configuration is treated separately, and in the other canonical sets of configurations are handled together. The latter method is theoretically preferred, although the format leads to a reduction in the number of configurations selected at a given energy threshold. Documentation for the Munich system is at present incomplete.
The MUNICH-CI program is being used, in conjunction with Dr D T Clarke and co-workers (Durham), in a theoretical investigation of the satellite spectra associated with core ionization for a series of small molecular systems. Dr Clarke is also eager to use the package in investigations of the electronic spectra of radical cations and calculations in the ethylene cation are underway.
During the last quarter two meetings of the Project 1 Working Group have been held. The first session, at State House, on October 16, reviewed the first years progress of Project 1. Satisfaction was expressed with the work done on the MUNICH-CI program, and arrangements made to convene a further meeting at ACD to draw up plans for future work. This took place on 4 December, and was open to other workers with an interest in correlation theory who to date had not been involved in the affairs of the Working Group. About 20 people attended and, after lengthy discussions, the following programme of work arrived at:
An ATMOL3 version of the ATMOL2 graphical analysis program is in the course of preparation. This program is large, and lacking in elegance of structure, because, of necessity, its construction involved the 'stitching up' of numerous routines and programs from a wide variety of sources. Hence its conversion to ATMOL3 form represents a considerable amount of work, during the completion of which we are also taking the opportunity to tidy up this system. Our aims are therefore 4-fold:
An improved version of the ATMOL2 file allocator is now available in ATMOL3.
Work is still proceeding at a slow but even rate on this part of the system, incorporating refinements and extra facilities as seem desirable.
Much of the work on ATMOL3 has been devoted to providing a write-up. It transpires that parts of the system now possess a write-up, although they are not yet coded. This method of work may seem eccentric, but has proved useful in that work may now proceed on those parts of the system with a frozen design and specification. We hope to complete the write-up plus associated coding (to bring the program's into line with the write-ups where necessary) in the next quarter, and during that period we do not expect to commence work on any new aspects.
Seventeen data sets were collected this quarter. There were no scanner maintenance problems and the program proved fully reliable. A new version of the Weissenberg program has been prepared. This uses the same dynamic core store allocation system as the precession program, and effectively combines the old 100 and 50 micron programs. The least squares section is identical to that used by the precession program.
The new version has now been successfully tested on two sets of photographs from small molecules and some more protein films. The 1906A program to which the Alpha 16 output is sent has been completed and tested.
The modifications for sampling scans have now been finished, and the work on improving this program is now essentially complete.
The X-Y recorder attachment for the Alpha-16 arrived this quarter and was installed and tested without difficulty. A considerable amount of programming effort has been put into providing software for this machine. The aim is to produce sets of routines which enhance the performance of the general scanner program. These fully test the plotter's capabilities, and have been checked out satisfactorily. Work on the more important routines for the X-ray film programs is well underway.
A new version of this program has been received and implemented satisfactorily on the 1906A. Copies have been sent to four other ICL machines. We have continued to use the program to monitor the accuracy of scanner data sets, and to clear up crystallographic problems that arise. Two new structures have been solved for the crystallographers concerned, using densitometer output.
There have been no problems with the system this quarter. As an exercise, an organic nickel perchlorate compound has been solved and fully refined using XRAY, from a diffractometer data set supplied by a user.
A successful demonstration of this program was made to the Council of the SRC in State House, using a dial-up line to the Oxford PDP10. The Data Compilation Committee have authorised a trial of the system at Manchester University using a link to Oxford, while work continues here on the task of implementing the file inversion routines. A preliminary attempt on the 1906A demonstrated the feasibility of producing inverted files that could be read by the PDP 10 and were correctly formatted. This work has now been transferred to the 360/195 (in anticipation of any moves to Daresbury) and is progressing satisfactorily.
The quarter began with a two-week visit to the Institute von Laue-Langevin in Grenoble for discussions concerning neutron scattering experiments and computer simulation methods. A great deal was talked about but so far nothing concrete has emerged from the visit, although the pilot project with Dr Lovesey of the ILL is progressing well.
The liquid crystal work is now being consolidated, with a number of papers being written. A film illustrating the order disorder phase transition in a model two-dimensional liquid crystal has also been made with the help of Simon Backer. This is to be shown at a Royal Institution discourse.
Professor Beeby and Mr Quirke of Leicester are attempting to calculate the Placek correction for neutron scattering from water using the molecular dynamics data of Rahman. There have been some snags partly with knowing how the data is written to the tape and some programming difficulties. At long last this seems to he working.
Finally, the project in collaboration with Professor March and Dr Grout of ICST has officially begun and is now proceeding fairly smoothly. We are trying to find out how pyroelectric materials work. Such materials are important as infra-red detectors.
After a great deal more chasing, IBM finally produced and installed the parts for the new 2741 terminal, except the correct golf-ball typing head, which still has not yet arrived. Otherwise, the new terminal is fully functional. The long delays in getting it going have dampened the initial enthusiasm of the secretarial staff, but Peter Smith has used it with some success in this quarter.
A talk was given to a meeting of the subgroup on the subject of Fortran and optimising compilers. A joint contribution is also to be made to a paper to be published in the Computer Journal discussing the proposed new ANSI Fortran standard. This is almost complete.
This work is continuing, in collaboration with A M Binnie of Cambridge, using the IBM 360/195 at Rutherford. The main problem to be overcome at the moment is the lack of convergence of the iteration on the free surface position. There appears to be a strong dependence on the initial position of the free surface, and there are constraints which cannot be built in to the iteration but which have to be satisfied (eg the water level should not become lower than ground level) which make life difficult. Under-relaxation of the free-surface iteration is being used to avoid oscillations which can produce unphysical results. The equation is highly non-linear so normal theory does not apply. All suggestions gratefully received!
Mr Binnie has requested further assistance with a similar problem, to be solved after the present work has been completed.
From time to time advice has been given to research workers from the universities and polytechnics on some aspect of numerical analysis or computation relevant to their problem. The latest request for assistance comes from the Rutherford Laboratory, asking for advice on the solution of a 2D diffusion problem over an awkward domain. This is currently being looked at to see whether one of the standard FE packages will cope, or whether a simple program could be written for this case. The equation itself is standard and should present no difficulties.
Due to the unavailability of the PDP15, work on a truly interactive system for FE packages has had to be postponed. However, a system has been devised, written and currently being tested that will allow a primitive form of interaction with the FE Package PAFEC through a MOP terminal on the 1906A.
Facilities allow regeneration of the main programs so that the program may run with totally different sets of data during the same job. This, together with the capability of viewing and erasure of drawings on the FR80 spool (via Spooljob macro), should make a valuable enhancement to the current version of PAFEC.
On Thursday 30 October SRC Engjneering Board members visited the Laboratory prior to their meeting at Cosener's House. Most of the demonstrations of interactive computing and graphics were undertaken by Basic Software Group. However, one item was included from Applications Software Group - the Finite Element film. Although in a state of complete unreadiness, and in spite of a Heath-Robinson contrivance to provide sound at more-or-less the same time as the picture, the visitors seemed to like the film and made favourable comments to those of us present.
A lot of preparation went into this visit, but viewed from the Public Relation angle it paid dividends; presenting a picture of an efficient and productive, and highly motivated Atlas staff.
Final touching-up is being done. This quarter has seen the "marrying" of the picture with the sound-track, and the advent of full colour.
It has been viewed not only by the Engineering Board (see elsewhere in this Report) but also by various visitors to the Laboratory, and at two conferences (Farnborough, with Paul Nelson, and Leicester, with Irene Buchanan). It received wide acclaim from these "outside" viewings and very many queries asking for copies for use in under-graduate and graduate courses all over the country.
Unfortunately we have had to stall these queries for the time being as no policy has yet emerged about the Film Library due to be set up to administer loans of SRC films. We also have had no directive concerning the sale of copies of the film. A brochure has been prepared in readiness for such a policy being approved, which can be printed within a month of the order being given, for distribution to all likely outlets eg university engineering departments.
Some thought has been given to the topic of software for the facility. Visits will be made in the next few weeks to a number of universities to see their machines and software systems.
The link from ROE to the 195 via ACD is now working and has been used regularly on a substantial scale since early December by Dr Thompson and Dr Morgan to run jobs from ROE. It was first brought into service in November and after some teething troubles with the contention system at ACD which controls access to the 195 (troubles which seem to have been associated exclusively with the use of the contention system by another RJE) it has been in regular use since early December.
The reprocessing of the S68 data started in November. As in the first run through the attitude refinement and signal identification program is run first for all orbits on one data tape; the parameter estimates for the attitude refinement solution are written to a file, the 'pass list', which is copied into a second file on the data tape; the estimates are also written to lineprinter with a concise summary of the signals observed and successful star matches. This information is inspected and checked. Then the main program is run in a mode which reads in the parameters from the 'pass list' to avoid repeating the attitude calculations, and then writes the raw data from the areas where star-like signals have been detected into various output files with information about their position and identification if the latter has been successful.
Data from extended objects is also extracted. The activity in the reprocessing has so far been confined to running the attitude refinement etc program on a number of tapes and thus building up a bank of tapes ready to go through the main program; which will be started in 1976.
The Problem Program Efficiency (PPE) was run on the attitude refinement program and showed a gratifying improvement in efficiency since it was last tried, in the early days of the first processing: then nearly 50% of the CPU time was spent in the FORTRAN I/O routines. Much of the improvement is due to the cumulative effect of many minor coding and structural changes: certainly each data tape is being processed in noticeably less CPU time and more information is being extracted from the data. Further improvements have since been made in areas which the PPE run showed to be heavily used.
David Carnochan of UCL had submitted an application to use the 1906A and the Tektronix terminal here: to display his data on 'blue objects', and to delete bad spectra. The attraction of the !906A fur this purpose is the much larger files accessible from MOP compared with ELECTRIC. The application is being considered; it may be possible to do what he wants on the 195 after all, by plotting the spectra on fiche, noting which are bad, and running a batch job to delete the bad ones.
The Belgians have after all continued to participate fully in the project. They have produced a lineprinter listing of the ultra-violet bright star catalogue which is their chief scientific responsibility, and the format of the published version is now being decided.
In the last three months Elizabeth Gill has been much involved in the preparations for the FILEMATCH workshop, with one or two other members of the team. Besides the FILEMATCH analysis programs and the G-EXEC interface, enhancements were necessary to the Executive Controller in G-EXEC to allow internal workfiles to be generated for given processes without the user's knowledge. The team also made unwillingly a passing acquaintance with the vagaries of COBOL when allied with 360 JCL - some troubles arose with the interface between SAFRAS and FILEMATCH and were eventually "patched" by the local team.
In the event the workshop went off well enough despite a disaster or two beforehand, (never arrange anything during an eclipse of the moon!!) - and some guarded enthusiasm was generated for FILEMATCH. Several helpful working contacts were made, particularly with the Office of Resource Analysis in Virginia (part of the USGS) and with more branches of the Geological Survey in Canada than hitherto. At the meeting in Paris immediately following the Workshop, plans were made for a pilot project of data exchange on two-mica granites, involving several centres in Europe and one or two in North America.
There are still no definite plans for the long-term future of G-EXEC as a complete project, but interest remains high in geological surveys and universities in various places in England and abroad for the distribution of the system. Development continues in response to the IGS workload, including a time sharing front-end system written hy Steve Henley in Edinburgh for the PDP 11-45 there. This is intended to include file-building and editing facilities as well as a simplified interactive Controller, and is at present under test. No use has yet been made of the UNIVAC version of the system at Wallingford, which was required by IGS primarily for its monthly work return project. The programmer required for this work is expected to arrive in January - meanwhile the system is kept ticking over by occasional test runs.
All commands other than GPLOT are now incorporated in the binary program resident in :SUBLIB. This may be accessed by the macro BOON2 held in :MACROS. The program requires 95K of store and has had to be overlaid with the unpacking and execution stages in two overlay units in order to mimimise lower data storage.
Problems with lower data also made it necessary to run the graphics routines as separate programs instigated by the BOON2 macro.
The GPLOT command has been delayed by modifications that became necessary on changing from the SD4020 to the FR80 plotter. These corrections have now been made and final tests on the plotting routines are being run.
The BOON Mark 2 manual has been finalised and approximately 50 copies have been distributed to possible users. The BOON Mark 2 system will be run only at Atlas for the time being. It will be distributed to other sites when the system appears to be bug-free.
This is substantially complete and about to he released to selected users on a trial basis. No further work is likely to be done except for maintenance as and when required.
| Discussions with staff at ILL, Grenoble, on possible NBRU collaboration, 2/3 October 1975 | B Stokoe, J W E Lewis |
| Meeting House, Project l Working Group meeting held at State House on Thursday 16 October 1975 | B Stokoe, V R Saunders, M F Guest, W R Rodwell |
| S2/68 Joint Executive Committee held at UCL on 20/21 October 1975 | B Stokoe, P F Smith |
| ALPS Meeting at ACD on 23 October 1975 | B Stokoe, P Kent, A J H Walter |
| Linguistic computing seminar in Oriental Inst, Oxford 30 October 1975 | B Stokoe |
| BOON Meeting at ACD held on 18 November 1975 | B Stokoe, P Kent |
| ACL/COGEODATA Workshop held at ACD on 21/22 November 1975 | B Stokoe, E M Gill, P G Sutterlin |
| COGEODATA CONFERENCE held in Paris on 24-26 November 1975 | B Stokoe, E M Gill, P G Sutterlin |
| Discussions at IGS on 17 December 1975 | B Stokoe |
| Data Compilation Committee meeting at State House on 18 December 1975 | B Stokoe |
| 4th Meeting of the Meeting. House Project 1 Working Group - 4 December 1976 | V R Saunders, M F Guest, W R Rodwell |
| NUMAC, Newcastle-on-Tyne to mount ATMOL at the request of Dr D T Clarke, 3-4 November 197.5 | V R Saunders |
| ICI Blackley, Manchester, to mount ATMOL at the request of Dr W C McCrodt, 14 November 1975 | V R Saunders |
| High-Power Lasers Discussion Meeting, Royal Society, London, 17 October 1975 | J E Crow |
| EPS High-Power Lasers Study Conference, RL 17-19 December 1975 | J E Crow |
| Meeting on Text Processing held at ACD on 17 November 1975 (with members of BSG) | P F Smith |
| Discussions with Dr O Kennard and her group at Cambridge, 30 October 1975 | P A Machin, M Elder |
| Demonstration of the CSSR System at Council meeting, 19 November 1975 | P A Machin |
| Conference on Simulation, held at Univ of Aston in Birmingham, 8 October 1975 | P Kent |
| Discussions at Imperial College on 29 October 1975 | J W E Lewis |
| Discussions at Univ of Southampton on 30 October 1975 | J W E Lewis |
| Discussions at Royal Hollaway College with Prof K Singer on 6 November 1975 | J W E Lewis |
| Discussions with Prof Rawlinson at Univ of Oxford on 14 November 1975 | J W E Lewis |
| Discussions with members of Prof March's group at Imperial College on 4 December 1975 | J W E Lewis |
| IGCP Working Group considering project for helping third world handle energy and mineral resource data in computerised form (Paris) 27-28 November 1975 | E M Gill |
| Special meeting to consider database management within NERC (Alhambra House) held on 13 November 1975 | E M Gill |
| Workshop on database management in universities held at Univ of Liverpool on 16. December 1975 | E M Gill |
| Summary of Progress in Project 1 of the Meeting House for period Oct 74-Sept 75 (MH/WG1/P4) 10 Oct 75 | V R Saunders |
| MUNICH-CI, A Progress Report (MH/WG1/P3) 10 October 1975. | M F Guest, W R Rodwell |
| MOLECULE-CI, A Progress Report; (MH/WG1/P2) 10 October 1976 | M F Guest, V R Saunders |
| Minutes of the Third Meeting of the Project 1 Working Group (MH/WG1/M3) 27 October i975 | V R Saunders, M F Guest |
| Coupled Pair Many Electron Theories (MH/WG1/P7) | V R Saunders, M A Robb (QE College) |
| Notes of and decisions taken at; the fourth meeting of Project 1 Working Group (MH/WG1/M4) | W R Rodwell |
| BOON Mark II, A User's Guide; dated October 1975 | P Kent, Cranfield |
| Talk on Geological Retrieval Systems to London Geological Society. 4 December 1975 | B Stokoe |
| Talk on FILEMATCH to London Geological Society 4 December 1975 | E M Gill |
| Report on visit to Paris to attend COGEODATA Conference, December 1975 | B Stokoe |
